Search code examples
jdbcdatabricksazure-databricksdatabricks-sql

How to get the cluster's JDBC/ODBC parameters programmatically?

Databricks documentation shows how get the cluster's hostname, port, HTTP path, and JDBC URL parameters from the JDBC/ODBC tab in the UI. See image:

Databricks's JDBC/ODBC Tab
(source: databricks.com)

Is there a way to get the same information programmatically? I mean using the Databricks API or Databricks CLI. I am particularly interested in the HTTP path which contains the Workspace Id.

Solution

  • You can use the Get operation of the SQL Analytics REST API (maybe together with List) - it returns the JDBC connection string as a part of response (jdbc_url field):

    {
  "id": "123456790abcdef",
  "name": "My SQL endpoint",
  "cluster_size": "Medium",
  "min_num_clusters": 1,
  "max_num_clusters": 10,
  "auto_stop_mins": 30,
  "num_clusters": 5,
  "num_active_sessions": 30,
  "state": "RUNNING",
  "creator_name": "[email protected]",
  "jdbc_url":"jdbc:spark://<databricks-instance>:443/default;transportMode=http;ssl=1;AuthMech=3;httpPath=/sql/protocolv1/o/0123456790abcdef;",
  "odbc_params": {
    "host": "<databricks-instance>",
    "path": "/sql/protocolv1/o/0/123456790abcdef",
    "protocol": "https",
    "port": 443
  }
}

    HTTP Path is also there, as path part of the odbc_params object.